Efforts to describe cephalometric data more efficiently involve abstracting closed curves through the use of a new geometry of form. The essence of this approach involves recoding cephalometric information in a way which capitalizes on the redundancy of biological growth processes. In this way it appears that fewer degrees of freedom will be required to duplicate the shape of a given structure with a fixed level of error.